Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Update Verizon Media copyright notices. | gjoranv | 2021-10-07 | 4 | -4/+4 |
| | |||||
* | Update 2018 copyright notices. | gjoranv | 2021-10-07 | 1 | -1/+1 |
| | |||||
* | Update 2017 copyright notices. | gjoranv | 2021-10-07 | 184 | -184/+184 |
| | |||||
* | Encapsulate in a context | Jon Bratseth | 2021-10-01 | 2 | -4/+6 |
| | |||||
* | Pass destination | Jon Bratseth | 2021-09-30 | 10 | -14/+72 |
| | | | | | This allows embedders to switch on it to enable bucket testing and similar. | ||||
* | encode -> embed | Jon Bratseth | 2021-09-28 | 11 | -62/+53 |
| | |||||
* | Set statement output explicitly | Jon Bratseth | 2021-09-26 | 5 | -0/+41 |
| | |||||
* | Remove outputType (not needed after all) | Jon Bratseth | 2021-09-19 | 2 | -25/+8 |
| | |||||
* | Add 'encode' expression | Jon Bratseth | 2021-09-19 | 19 | -30/+206 |
| | |||||
* | Set output type | Jon Bratseth | 2021-09-17 | 2 | -9/+24 |
| | |||||
* | Non-functional changes only | Jon Bratseth | 2021-09-17 | 54 | -100/+93 |
| | |||||
* | Non-functional changes only | Jon Bratseth | 2021-09-17 | 53 | -152/+162 |
| | |||||
* | No functional changes | Jon Bratseth | 2021-09-17 | 19 | -12/+32 |
| | |||||
* | we want to compare Linguistics objects for equivalence | Arne Juul | 2021-08-04 | 1 | -0/+3 |
| | |||||
* | avoid depending on Linguistics equals() | Arne Juul | 2021-08-03 | 1 | -1/+7 |
| | |||||
* | do not depend on Linguistics equals() | Arne Juul | 2021-08-03 | 1 | -1/+1 |
| | | | | | | * a Linguistics object typically has no state and does not implement the semantics we want for equals() method. * alternately, we could remove this check. | ||||
* | don't call accentDrop at all for empty input | Arne Juul | 2021-07-16 | 2 | -2/+28 |
| | |||||
* | try to trap spurious failure | Arne Juul | 2021-07-13 | 2 | -0/+65 |
| | | | | | | * we have seen spurious failures when verifying output from accent dropping; so far nothing reproducible, so add some extra logging and retry once if it happens (in case it's some kind of race-condition glitch). | ||||
* | Add a test | Jon Bratseth | 2021-07-02 | 2 | -8/+6 |
| | |||||
* | No functional changes | Jon Bratseth | 2021-06-28 | 1 | -11/+10 |
| | |||||
* | Require replacements to be applied during tokenization | Jon Bratseth | 2021-06-15 | 2 | -28/+13 |
| | |||||
* | Support literal 'true' and 'false' | Jon Bratseth | 2021-04-30 | 9 | -3/+147 |
| | |||||
* | Support to_bool | Jon Bratseth | 2021-04-30 | 7 | -2/+152 |
| | |||||
* | Avoid changing the input string field value by cloning it before doing changes. | Geir Storli | 2021-01-07 | 1 | -5/+8 |
| | | | | | This is needed as the input can be used by other expressions as well. The same cloning is done in ExactExpression and TokenizeExpression. | ||||
* | Merge pull request #13709 from ↵ | Jon Bratseth | 2020-06-26 | 1 | -2/+2 |
|\ | | | | | | | | | vespa-engine/bratseth/surrogate-aware-gram-splitting Surrogate aware gram splitting | ||||
| * | Surrogate aware gram splitting | Jon Bratseth | 2020-06-25 | 1 | -2/+2 |
| | | |||||
* | | Handle an existing spantree in exactmatch | Jon Bratseth | 2020-06-24 | 2 | -6/+13 |
|/ | | | | | This may happen if a field which is indexed is used as an input for another field indexed as exact match. | ||||
* | Disallow unbound tensor dimensions in document fields | Jon Bratseth | 2020-06-08 | 2 | -0/+3 |
| | |||||
* | Remove use of apache commons libraries in indexinglanguage | Bjørn Christian Seime | 2020-01-03 | 4 | -11/+7 |
| | |||||
* | Preserve array updates with element index matching in indexing docproc | Tor Brede Vekterli | 2019-12-04 | 4 | -5/+102 |
| | | | | | | The resulting `MapValueUpdate` would previously be constructed with the wrong type for the index, causing an exception to be thrown and for the update to fail entirely. | ||||
* | Preserve AssignValueUpdates to struct fields in indexing docproc | Tor Brede Vekterli | 2019-12-02 | 2 | -6/+28 |
| | | | | | | | Would previously be rewritten as MapValueUpdates for unknown reasons. This wouldn't actually work, as an exception would be thrown during serialization when sanity checking code figured out that the MapValue update was attempted used for a non-array/wset field data type. | ||||
* | doc: -> id: | Henning Baldersheim | 2019-08-16 | 3 | -11/+11 |
| | |||||
* | doc: -> id: | Henning Baldersheim | 2019-08-16 | 6 | -21/+21 |
| | |||||
* | Merge pull request #9969 from vespa-engine/geirst/more-robust-ngram-expression | Arne H Juul | 2019-07-08 | 2 | -0/+20 |
|\ | | | | | Ensure that NGramExpression can be executed multiple times, where spa… | ||||
| * | Ensure that NGramExpression can be executed multiple times, where span tree ↵ | Geir Storli | 2019-07-05 | 2 | -0/+20 |
| | | | | | | | | | | | | | | from first execution is used. This can happen when we have multiple UpdateAdapters (e.g. regular updates + field path updates) and then all scripts are executed per adapter in Expression.execute(). | ||||
* | | Nonfunctional changes only | Jon Bratseth | 2019-07-03 | 1 | -0/+1 |
|/ | |||||
* | Assert that TensorRemoveUpdate passes through indexinglanguage | Lester Solbakken | 2019-02-15 | 3 | -7/+61 |
| | |||||
* | Test that TensorModifyUpdate and TensorAddUpdate pass through unmodified. | Geir Storli | 2019-02-14 | 2 | -2/+62 |
| | |||||
* | Pass TensorAddUpdate right through indexing docproc. | Geir Storli | 2019-02-11 | 2 | -1/+6 |
| | |||||
* | Pass TensorModifyUpdate right through indexing docproc. | Geir Storli | 2019-02-01 | 2 | -0/+5 |
| | |||||
* | Make SimpleLinguistics simple again | Jon Bratseth | 2019-01-21 | 3 | -3/+3 |
| | | | | | - Remove SimpleLinguistics config and optional use of Optimaize - Add Optimaize to OpennlpLinguistics; on by default and config to disable | ||||
* | 6-SNAPSHOT -> 7-SNAPSHOT | Arnstein Ressem | 2019-01-21 | 1 | -2/+2 |
| | |||||
* | Update to latest ph-javacc-maven-plugin. | gjoranv | 2018-12-06 | 1 | -6/+7 |
| | | | | | | | | | - Update FastCharStream method names to follow changes in generated CharStream class. - Update expected test output due to improvements in generated code. - Improve ability to debug tests by more clearly printing out the diff between expected and actual output. | ||||
* | Manage ph-javacc-maven-plugin config in parent pom. | gjoranv | 2018-12-06 | 1 | -9/+0 |
| | | | | + Remove unused management for old javacc plugin. | ||||
* | Deprecated methods and add OptimaizeDetector | Jon Bratseth | 2018-11-01 | 3 | -0/+3 |
| | |||||
* | Use hashmap for faster access. | Henning Baldersheim | 2018-10-25 | 5 | -37/+37 |
| | |||||
* | Use maven-shade-plugin to assemble fat jar | Bjørn Christian Seime | 2018-10-19 | 1 | -7/+15 |
| | | | | | A bug in maven-assemble-plugin makes it very slow to assemble jar containing BouncyCastle | ||||
* | Make inputDataType final to ensure complete at construction time. | Henning Baldersheim | 2018-10-17 | 19 | -78/+88 |
| | |||||
* | Move all input type verification to expression construction time. | Henning Baldersheim | 2018-10-17 | 61 | -361/+239 |
| | |||||
* | Compute the type generated by the Expression once. | Henning Baldersheim | 2018-10-16 | 1 | -0/+16 |
| |